110 ALASKAN GLACIER STUDIES southeast of the glacier is proved by morainic margins so recently abandoned as to be free from vegetation. Gilbert had previously shown that in 1899 *• Hubbard Glacier had retreated one or two hundred feet since Russell made his studies in 1890 and 1891, as viewed from Osier Island and slightly since 1895 as viewed from Haenke Island. The retreat was even greater between 1899 and 1901, for Gilbert states that the 1901 pictures by the Fish Commission party show that "at a point where a prominent moraine makes the comparison somewhat definite, the ice cliff appears to have stood 700 to 1000 feet farther back than in 1899. The cliff was also shortened at each end by the enlargement of the marginal belts of stagnant ice." Between 1899 and 1905 the northwest part of Hubbard Glacier advanced, as shown by comparison of our photographs with those from Gilbert's photographic site above Osier Island. The west margin was farther out in 1905 than when photographed from Osier Island by Russell in 1891. It was several hundred yards farther out than when photographed from Haenke Island by Russell in 1891 and by Gilbert in 1899. The bend in the medial moraine of the northwest arm, as seen from Haenke Island and from the mountain side above Osier Island, was a quarter of a mile farther southeast in 1905 then in 1891, and several hundred yards farther in 1905 than in 1899. The exact site of Gilbert's 1899 pictures of Hubbard Glacier from the coast between Variegated and Hubbard Glaciers could not be found in 1905 because altered by wave work; but from approximately the same site, that is with the same points in line on the mountains in the background, it was perfectly clear that the southeast side of Hubbard Glacier had retreated between 1899 and 1905, probably several hundred yards. This advance of the northwest part of Hubbard Glacier and retreat of the southeast side between 1899 and 1905 we interpret as an indication of an advance of the weaker northwest arm under the impulse of the earthquake shaking, the strong north arm retreating relatively because of ice lost during the severe shaking of the 1899 earthquakes, as proved by the direct testimony of prospectors who were encamped beside it in 1899 * and by the Fish Commission pictures from which Gilbert proved a retreat of 700 to 1000 feet between 1899 and 1901, as already stated. Condition in 1906. In this year we thought that there was a more active discharge of icebergs from Hubbard Glacier front than in the previous season, and that there was a greater quantity of ice floating in Disenchantment Bay. But such evidence is of little value, being mere impression. Comparison with the photographs of 1905, however, gives evidence of a slight advance in the northwestern portion of the ice front, but nothing to compare with the advance of other glaciers in the region, including the smaller Variegated and Haenke Glaciers nearby. We made a visit to the southeastern margin of the Hubbard and found the condition there unchanged, so that the description given in previous pages applies as fully in 1906 as to the preceding year. Advance in 1909. In the summer of 1909 evidence was observed of a very distinct forward movement (PI. XLVIII), though it was apparently only the beginning. This i Gilbert, G. K, Glaciers and Glaciation, Harrunan Alaska Expedition, Vol. 8,1904, pp. 63-66. i Tarr, R. 3. and Martin. Lawrence, Recent Changes of Level in Alaska, Bull. Geol. Soc. Amer., Vol. XVII, 1906, p. 31; Martin, Lawrence, The Alaskan Earthquakes of 1899, Bull. Geol. Soc. Amer., Vol. 21,1910, p. 359; Tarr, R. S. and Martin, Lawrence, The Earthquakes at Yakutat Bay, Alaska, in September, 1899, Professional Paper No. 69, TJ. S. Geol. Survey, 1912, pp. 15-17.